TCP over High Speed Variable Capacity Links: A Simulation Study for Bandwidth Allocation
|
|
- Sharleen Wood
- 8 years ago
- Views:
Transcription
1 TCP over High Speed Variable Capacity Links: A Simulation Study for Bandwidth Allocation Henrik Abrahamsson, Olof Hagsand and Ian Marsh SICS AB, Kista S- 9, Sweden {henrik, ianm}@sics.se Dynarc AB, Kista S-, Sweden hagsand@dynarc.se Abstract. New optical network technologies provide opportunities for fast, controllable bandwidth management. These technologies can now explicitly provide resources to data paths, creating demand driven bandwidth reservation across networks where an applications bandwidth needs can be meet almost exactly. Dynamic synchronous Transfer Mode (DTM) is a gigabit network technology that provides channels with dynamically adjustable capacity. TCP is a reliable end-to-end transport protocol that adapts its rate to the available capacity. Both TCP and the DTM bandwidth can react to changes in the network load, creating a complex system with inter-dependent feedback mechanisms. The contribution of this work is an assessment of a bandwidth allocation scheme for TCP flows on variable capacity technologies. We have created a simulation environment using ns- and our results indicate that the allocation of bandwidth maximises TCP throughput for most flows, thus saving valuable capacity when compared to a scheme such as link over-provisioning. We highlight one situation where the allocation scheme might have some deficiencies against the static reservation of resources, and describe its causes. This type of situation warrants further investigation to understand how the algorithm can be modified to achieve performance similar to that of the fixed bandwidth case. Keywords: TCP, DTM, rate control, rate adaption Introduction Reliable transfer of data across the Internet has become an important need. TCP [Pos8] is the predominant protocol for data transfer on the Internet as it offers a reliable end-to-end byte stream transport service. Emerging optical networking technologies provide fast, cheap and variable capacity bandwidth links to be setup in milliseconds allowing data-driven virtual circuits to be created when needed. One example of an application that could use such a service is the backup of critical data. Exact allocation of bandwidth to TCP flows would alleviate complex traffic engineering problems such as provisioning and dimensioning. Allocating bandwidth to TCP is a complex problem; the TCP congestion control mechanism
2 plus network dynamics can make exact allocation for TCP data flows difficult. Te contribution of this paper is the performance evaluation of an estimation algorithm, which measures the rate of TCP flows and allocates capacity on a DTM network. Dynamic Synchronous Transfer Mode [GHP9] [BHL + 9] is a gigabit ring based networking technology that can dynamically adjust its bandwidth. DTM offers a channel abstraction, where a channel consists of a number of slots. The number of slots allocated to a channel determines its bandwidth. The slots can be allocated statically by pre-configured parameters, or dynamically adjusted to the needs of an application. In DTM it is possible to allocate a channel to a specific TCP connection, or to multiplex several TCP connections over the same channel. We mostly investigated cases where each TCP connection is assigned to a separate channel, but show one case in which two TCP connections compete for a single channel. TCP uses an end-to-end congestion control mechanism to find the optimal bandwidth at which to send data. In order to get good throughput with TCP operating over a technology such as DTM, it is important to understand the dynamic behaviour of the two schemes, especially when evaluating a bandwidth allocation strategy. TCP is capable of adjusting its rate whilst DTM is capable of changing its capacity. In dynamically interacting systems, it is possible to create unwanted oscillations resulting in under allocation or over allocation of bandwidth to TCP flows. In order to evaluate the performance of the DTM bandwidth allocator, we have implemented the algorithm in the network simulator ns-. We have performed a number of simulations that include single and multiple TCP flows, links with varying delay characteristics, different buffer sizes, plus TCP Reno and Tahoe variants. Section outlines DTM and our estimation algorithm, simulation experiments are given in Section, related work follows in Section, and finally conclusions and a discussion are given in Section. Dynamic Synchronous Transfer Mode DTM uses a TDM scheme where time slots are divided into control and data slots. The control slots are statically allocated to a node and are used for signalling. Every node has at least one control slot allocated that corresponds to kbps of signalling capacity. The data slots are used for data transmission and each slot is always owned by a node. A node is only allowed to send in slots that it owns. The ownership of the slots is controlled by a distributed algorithm, where the nodes can request slots from other nodes. The algorithms for slot distribution between the nodes affect the network performance. Each slot contains bits and the slots are grouped in microsecond long cycles. The bit rate is determined by the number of slots in a cycle, so one slot corresponds to a bit rate of kbps. By allocating a different numbers of slots, the transmission rate for a channel can be changed in steps of kbps.
3 . TCP Rate Estimation and DTM Capacity Allocation TCP s rate is simply estimated as the number of incoming bytes per second. The algorithm which is presented next calculates the rate by dividing the number of bytes by the time elapsed. The rate of each flow is calculated ten times per second, i.e. every ms. This value has been chosen as a compromise between good measurement granularity and processing overhead. Actual slot allocation or changes are done only once every second, this is slightly coarser due to the overhead of nodes potentially having to negotiate slots. We now describe the TCP bandwidth estimator. Figure shows the algorithm used to estimate the rate of a given flow. As stated, every ms the estimator measures the rate new in bits per second and compares it with the previous value, current. A delta of the difference is reduced by DTM SHIFT in the algorithm. Note this delta is simply shifted, keeping the complexity of the calculation to a minimum. In this case it is three, so the current value is changed by one eighth towards the recently measured flow value, as shown in the first half of the algorithm. This shift effectively determines how aggressively TCP s rate can be tracked. Finally the units are changed from bits per second to slots per second by dividing the rate by the channel bandwidth and assigning this value to the variable curr slot. dtm calc bw ( new ) { DTM SHIFT = MARGIN =.7 CORRIDOR = /* first half - Move last estimate closer */ diff = new - current if ( diff < ) { diff = (-diff) DTM SHIFT current = current - diff // Decreasing } else { diff = diff DTM SHIFT current = current + diff // Increasing } curr slot = current / slot bw /* Second half - Last estimate within bounds? */ if ( curr slot > upper bound ) (curr slot < lower bound ) { dynbw = curr slot + MARGIN + ( CORRIDOR / ) /* only change bw once per sec */ change link bw (dynbw) } } Fig.. Algorithm for bandwidth estimation
4 The second half of the algorithm determines whether it is necessary to change the slot allocations. The current slot value is compared to upper and lower bounds before making any changes. An offset,.7 of a slot, MARGIN equivalent to 9 kbits, is added to the TCP throughput estimate so the DTM allocation will be a little over the estimated rate. Figure shows two plots using the topology shown in Figure, the leftmost plot is the actual measured bandwidth of a single TCP flow. The right plot shows the effect of adding MARGIN and measuring the rate in slots. If the allocation was based purely on this estimate it would under allocate bandwidth, causing TCP reduce its window because of congestion on the link. The rightmost graph is coarser due to the second granularity of the bandwidth changes. The plots illustrate how the estimation can be used to give TCP the bandwidth it needs and hence maximise throughput. One can also see in this figure that estimation starts after ms but a change is not applied to the offered bandwidth before the first second. Note also the y-axis in Figure b) is in slots per second and not bits per second as in the left figure. Additionally, a CORRIDOR is an amount the estimate is allowed to vary before slots are added or decreased for a channel. This is not visible in the plots but will be illustrated later. The purpose is to avoid small fluctuations causing unnecessary costly slot allocation changes. As mentioned, slot changes can be time consuming due to the distributed nature of DTM [AMMS98]..e+7 TCP rate DTM estimation rate TCP rate Allocated Slots throughput bits/sec e+7 8e+ e+ e+ e+ Slots/sec (a) Estimating the TCP Rate (b) DTM slot allocation Fig.. Measured Throughput and Slot Allocation
5 Sender / packet buffer.. TCP Estimator Receiver node ec ms node ec ms node DTM Link node ec /ms node Fig.. Simulation topology Simulation Tests This section presents simulation results that show how the DTM estimation algorithm adapts the offered bandwidth to TCP flows. Figure shows the topology we used for the following simulations. The Mbits per second link between nodes two and three is the bottleneck link. The link between nodes three and four is the DTM link with dynamically allocated bandwidth. Initially the DTM link is set to Mbits per second. This value was chosen simply for convenience, since simulating a Mbits per second link with large bandwidth flows is not feasible in a packet level simulator like ns-. The other two links also have a capacity of Mbits per second. A bulk transfer TCP Reno flow was setup between nodes one and five and the throughput measured at node three, in order to allocate bandwidth on the outgoing DTM link. In this first simulation the queue length in node was set to packets, figure shows the result. In congestion avoid (a) Congestion Window (b) Throughput (c) DTM allocation Fig.. Dynamically allocated bandwidth on a DTM link ( packet queue) ance the TCP flow increases the congestion window by the maximum segment size bytes each RTT seconds. However, the increase is not made each RTT. Instead TCP will increase MSS/congestion window bytes each time an ACK is
6 received. This means that after RTT seconds, the congestion window was increased by MSS bytes. This continues until the TCP flow has filled the buffer space at the bottleneck link, resulting in a packet drop. TCP Reno, using fast retransmit and fast recovery, then reduces the congestion window by half and continues with congestion avoidance. The congestion window, therefore, follows a sawtooth curve. If enough buffer space is available at the bottleneck link, the rate of the TCP flow, perceived after the second link, is not affected when the congestion window is reduced. This mechanism and result can be seen in left and middle plots of Figure. The rightmost plot shows the dynamically allocated bandwidth on the DTM link. It can be seen that TCP actually manages to get about one Megabit per second more on the DTM link due to the extra capacity allocated to the flow through the addition of MARGIN. It should be stated in a real deployment of TCP over DTM that this value is settable by network operators. Its affect can be tested in simulation environments such as this if necessary. Figure shows the results when the queue size at the bottleneck link is limited to ten packets. This could be the case if a static allocation over the DTM network has been setup. Now the rate of the TCP flow changes with the congestion window, but the changes are too small to affect the dynamic allocation of bandwidth. This is due to the corridor mentioned earlier to avoid small changes from incurring changes in the slot allocation scheme. Figure 7 8 (a) Congestion Window (b) Throughput (c) DTM allocation Fig.. Dynamically allocated bandwidth on a DTM link ( packet queue) shows the case in which the simulation with a small queue size and a ms link delay has been repeated using TCP Tahoe instead of TCP Reno. TCP Tahoe only relies on the retransmission timer and does not use fast retransmit. When a packet is dropped, the congestion window is set to one and slow-start is invoked. We can see that the allocation on the DTM link closely follows the sharp saw tooth behaviour of TCP Tahoe Figure c).
7 8 8 (a) Congestion Window (b) Throughput (c) DTM allocation Fig.. Tahoe TCP on a DTM link ( packet queue). Two Flows Per Channel and Small Router Buffers So far, we have shown cases where the dynamic allocation of bandwidth has allowed TCP to maximise its throughput. We illustrate one case next when the algorithm has weaknesses to allocate sufficient bandwidth to two TCP flows. In this scenario the fixed link case performs better. Figure 7 shows the topology Sender ec ms packets... TCP Rate Estimator Experiment ) Fixed Capacity Experiment ) Variable Capacity ec ms Receiver DTM Link node node node ec node node ec ec ec ms ms ms node node 7 Sender Receiver Fig. 7. Simulation topology with two flows that we used. It differs from previous simulations in that the flows have their own input buffer at node two but share a common output buffer in the same node. This buffer is also served ten times faster than in previous cases by the fact that the link feeding the DTM network was set to Mbits per second. In this case the queue length of a DTM link, node three, was limited to ten packets.
8 Figure 8 shows the results when the link capacity between nodes and was fixed at Mbits per second. We can see that both flows manage to reach their Mbits throughput, effectively sharing equally one DTM channel. If we now turn our attention to the same simulation but replace the static link between nodes three and four with a variable one the results are quite different. Figure 9 shows the dynamically allocated bandwidth on the DTM link. Neither of the flows manage to reach Mbits per second on their output links. In this case packets are being dropped in the output buffer of node three. This can be seen in the congestion windows of the two flows, they never manage to maintain the size of the static case, about segments. The problem in this case is the estimation algorithm should not decrease the estimation if packets are being dropped. The algorithm is symmetric, it increases or decreases depending on the measured rate. Additionally the effect of the short queue does not help, there is not sufficient pressure with a small queue to keep the rate up, with a larger buffer there is more pressure due to accumulated packets. Interestingly, the algorithm actually correctly allocates for the observed throughput, however does not maximise the TCP throughput. Related work Work on estimating and maximising TCP throughput for variable capacity links is relatively scarce. However comprehensive studies have been done related to the performance of TCP on ATM networks [Bon98] [MG9] [CLN9]. The main conclusions of the works are similar, the traffic classes of ATM are poorly suited to the bursty needs of TCP, due to the traffic contracts needed by ATM classes. The conclusion of [Bon98] is that the complexity of choosing traffic parameters for ABR is not in proportion to the benefits of carrying TCP/IP traffic. The CBR class is too simple for TCP, as only the peak rate is specified. Most of the DTM research in this area focuses on the distributed slot allocation for example [AMMS98]. Clark and Fang propose a framework for allocating bandwidth to different users during congestion [CF98]. The focus of the work is TCP bulk-data transfers. The authors attempt to keep TCP flows in congestion avoidance in the best case, and fast recovery phase in the worst case, by avoiding dropping several packets of the same flow in the same RTT. The conclusions of the given work are similar to those of [Bon98], that TCP connections can have difficulties to fill their alloted bandwidth. The work resembles ours in that they attempt to allocate bandwidth between different flows in a fair manner. It differs from ours in that we assume that the network can change its offered bandwidth and we focus on maximising TCP throughput, rather than trying to maintain a TCP state in the face of adverse network conditions. In addition, we allocate bandwidth to flows not only when the network is congested but also in normal situations as well. Sterbenz and Krishnan investigate TCP over Load-Reactive Links [KS]. They use a hysteresis control mechanism for capacity allocation. Buffer levels
9 8 (a) Window Flow (b) Rate Flow 8 (c) Window Flow (d) Rate Flow Fig. 8. Experiment static link: The senders do not drop packets at the ingress node and achieve their constrained link throughput Mbits per second.
10 8 (a) Window Flow (b) Rate Flow 8 8 (c) Window Flow (d) Rate Flow (e) Aggregated DTM rate Fig. 9. Experiment Dynamic Link: Allocated bandwidth on DTM link. Congestion window and measured rate for two TCP Reno flows with limited queue size at the DTM link.
11 are monitored (as in [Lun98]) and if the occupancy is greater than a threshold the capacity is increased and vice versa. This approach is not the same as ours, we measure the rate of incoming TCP flows at the router before the DTM link rather than the buffer level in the router at the outgoing DTM link. A single TCP flow is simulated and the authors state that the control parameters should be carefully chosen. Poor parameter choice can have the opposite effect, resulting in TCP not being able to operate. The work resembles ours in that a method is presented to react to network load and allocate bandwidth for TCP accordingly. It differs in that we measure the throughput of individual flows and allocate bandwidth from this measurements, where they use the buffer length as a measure of the load. Our system is less scalable, but more accurate, as we can ascertain exactly the bandwidth of the incoming TCP connections. Lundqvist evaluates different algorithms for bandwidth allocation for DTM channels transporting IP traffic [Lun98]. The algorithms were assessed with respect to throughput, delay and bandwidth changes per second. TCP rate adjustment is done by placing the incoming packets into a buffer and adding and removing slots if the level of the buffer exceeds continuously maintained threshold values. He concludes that adaptive strategies are recommended for TCP, however too frequent changes can be undesirable in a DTM network due to the processing cost. The main conclusion from this work is that the choice of algorithm can play a significant role in the performance. This work is similar to ours in that the goal is a slot allocation for TCP traffic over DTM. We also agree it is important to keep the computational complexity low and DTM bandwidth changes as infrequent as possible. It differs from ours in that we measure the rate of each TCP flow, whilst he looks at the outgoing buffer length as a sign to increase or decrease the number of slots. We look more into network scenarios such as different link delays, buffer lengths and use two different TCP types, TCP Reno and Tahoe. Conclusions We have analysed a complex problem, allocating bandwidth to a protocol that can adapt its rate. The benefits of guaranteeing throughput for an application using TCP can be very beneficial, in particular the cost savings when paying per unit of transmission. The goal was to investigate the behaviour of our bandwidth estimation scheme, its affect on TCP and on a network that can vary its capacity, in this case DTM. Our work however is not only limited to DTM technology, we can draw the same conclusions about TCP performance on any high speed network technology that offers variable capacity. We have written a simulation environment using ns-, and found that in almost all cases, TCP could be allocated a share of the channel identical to its measured throughput on a fixed network. We identified one scenario in which the algorithm could be improved, when packets are dropped at a router with a small buffer. In this situation the estimation algorithm should not reduce the offered bandwidth further, resulting in less offered bandwidth and further packet loss.
12 Instead it should allow TCP to find the new capacity available in the network. The combination of the small buffer size plus high speed input link aggravates this observed deficiency. In a simulation environment the parameter space is large. Due to space limitations we have only discussed a key subset of possible buffer sizes, link bandwidths, link delays and TCP variants. Further results, plus validation tests for using ns- in these kind of simulations, can be found in the technical report [AM]. Parameters that are worthy of further investigation include sampling times and estimation thresholds. Acknowledgements We would sincerely like to acknowledge the Computer and Network Architecture lab at SICS, and the Laboratory of Communication Networks at KTH University, Stockholm. References [AM] Henrik Abrahamsson and Ian Marsh. Dtmsim - dtm channel simulation in ns. Technical Report T:, SICS Swedish Institute of Computer Science, November. [AMMS98] Csaba Antal, József Molnár, Sándor Molnár, and Gabor Szabó. Performance study of distributed channel allocation techniques for a fast circuit switched network. Computer Communications, (7):97 9, November 998. [BHL + 9] Christer Bohm, Markus Hidell, Per Lindgren, Lars Ramfelt, and Peter Sjödin. Fast circuit switching for the next generation of high performance networks. IEEE Journal on Selected Areas in Communications, ():98, February 99. [Bon98] O. Bonaventure. Integration of ATM under TCP/IP to provide services with minimum guaranteed bandwidth. PhD thesis, University of Liege, 998. [CF98] D. Clark and W. Fang. Explicit allocation of best-effort packet delivery service. IEEE/ACM Transactions on Networking, (), August 998. [CLN9] E. Chan, V. Lee, and J. Ng. the performance of bandwidth allocation strategies for interconnecting atm and connectionless networks, 99. [GHP9] L. Gauffin, L. H. Hakansson, and B. Pehrson. Multi-gigabit networking based on DTM. Computer Networks and ISDN Systems, ():9, April 99. [KS] Rajesh Krishnan and James Sterbenz. Tcp over load-reactive links. In International Conference on Network Protocols (ICNP),. [Lun98] Henrik Lundqvist. Performance evaluation for IP over DTM. Master s thesis, Linköping University, Linköping, 998. [MG9] [Pos8] Kjersti Moldeklev and Per Gunningberg. How a large ATM MTU causes deadlocks in TCP data transfers. IEEE/ACM Transactions on Networking, ():9, 99. J. Postel. Transmission control protocol. Request for Comments 79, Internet Engineering Task Force, September 98.
Improving the Performance of TCP Using Window Adjustment Procedure and Bandwidth Estimation
Improving the Performance of TCP Using Window Adjustment Procedure and Bandwidth Estimation R.Navaneethakrishnan Assistant Professor (SG) Bharathiyar College of Engineering and Technology, Karaikal, India.
More informationAN IMPROVED SNOOP FOR TCP RENO AND TCP SACK IN WIRED-CUM- WIRELESS NETWORKS
AN IMPROVED SNOOP FOR TCP RENO AND TCP SACK IN WIRED-CUM- WIRELESS NETWORKS Srikanth Tiyyagura Department of Computer Science and Engineering JNTUA College of Engg., pulivendula, Andhra Pradesh, India.
More informationTCP over Multi-hop Wireless Networks * Overview of Transmission Control Protocol / Internet Protocol (TCP/IP) Internet Protocol (IP)
TCP over Multi-hop Wireless Networks * Overview of Transmission Control Protocol / Internet Protocol (TCP/IP) *Slides adapted from a talk given by Nitin Vaidya. Wireless Computing and Network Systems Page
More informationLecture 15: Congestion Control. CSE 123: Computer Networks Stefan Savage
Lecture 15: Congestion Control CSE 123: Computer Networks Stefan Savage Overview Yesterday: TCP & UDP overview Connection setup Flow control: resource exhaustion at end node Today: Congestion control Resource
More informationRobust Router Congestion Control Using Acceptance and Departure Rate Measures
Robust Router Congestion Control Using Acceptance and Departure Rate Measures Ganesh Gopalakrishnan a, Sneha Kasera b, Catherine Loader c, and Xin Wang b a {ganeshg@microsoft.com}, Microsoft Corporation,
More information17: Queue Management. Queuing. Mark Handley
17: Queue Management Mark Handley Queuing The primary purpose of a queue in an IP router is to smooth out bursty arrivals, so that the network utilization can be high. But queues add delay and cause jitter.
More informationPART II. OPS-based metro area networks
PART II OPS-based metro area networks Chapter 3 Introduction to the OPS-based metro area networks Some traffic estimates for the UK network over the next few years [39] indicate that when access is primarily
More informationTransport layer issues in ad hoc wireless networks Dmitrij Lagutin, dlagutin@cc.hut.fi
Transport layer issues in ad hoc wireless networks Dmitrij Lagutin, dlagutin@cc.hut.fi 1. Introduction Ad hoc wireless networks pose a big challenge for transport layer protocol and transport layer protocols
More informationTransport Layer Protocols
Transport Layer Protocols Version. Transport layer performs two main tasks for the application layer by using the network layer. It provides end to end communication between two applications, and implements
More informationCongestion Control Review. 15-441 Computer Networking. Resource Management Approaches. Traffic and Resource Management. What is congestion control?
Congestion Control Review What is congestion control? 15-441 Computer Networking What is the principle of TCP? Lecture 22 Queue Management and QoS 2 Traffic and Resource Management Resource Management
More informationInternational Journal of Scientific & Engineering Research, Volume 6, Issue 7, July-2015 1169 ISSN 2229-5518
International Journal of Scientific & Engineering Research, Volume 6, Issue 7, July-2015 1169 Comparison of TCP I-Vegas with TCP Vegas in Wired-cum-Wireless Network Nitin Jain & Dr. Neelam Srivastava Abstract
More information1. The subnet must prevent additional packets from entering the congested region until those already present can be processed.
Congestion Control When one part of the subnet (e.g. one or more routers in an area) becomes overloaded, congestion results. Because routers are receiving packets faster than they can forward them, one
More informationActive Queue Management (AQM) based Internet Congestion Control
Active Queue Management (AQM) based Internet Congestion Control October 1 2002 Seungwan Ryu (sryu@eng.buffalo.edu) PhD Student of IE Department University at Buffalo Contents Internet Congestion Control
More informationAnalyzing Marking Mod RED Active Queue Management Scheme on TCP Applications
212 International Conference on Information and Network Technology (ICINT 212) IPCSIT vol. 7 (212) (212) IACSIT Press, Singapore Analyzing Marking Active Queue Management Scheme on TCP Applications G.A.
More informationPerformance Analysis of AQM Schemes in Wired and Wireless Networks based on TCP flow
International Journal of Soft Computing and Engineering (IJSCE) Performance Analysis of AQM Schemes in Wired and Wireless Networks based on TCP flow Abdullah Al Masud, Hossain Md. Shamim, Amina Akhter
More informationLecture Objectives. Lecture 07 Mobile Networks: TCP in Wireless Networks. Agenda. TCP Flow Control. Flow Control Can Limit Throughput (1)
Lecture Objectives Wireless and Mobile Systems Design Lecture 07 Mobile Networks: TCP in Wireless Networks Describe TCP s flow control mechanism Describe operation of TCP Reno and TCP Vegas, including
More informationFinal for ECE374 05/06/13 Solution!!
1 Final for ECE374 05/06/13 Solution!! Instructions: Put your name and student number on each sheet of paper! The exam is closed book. You have 90 minutes to complete the exam. Be a smart exam taker -
More informationAn enhanced TCP mechanism Fast-TCP in IP networks with wireless links
Wireless Networks 6 (2000) 375 379 375 An enhanced TCP mechanism Fast-TCP in IP networks with wireless links Jian Ma a, Jussi Ruutu b and Jing Wu c a Nokia China R&D Center, No. 10, He Ping Li Dong Jie,
More informationQoS issues in Voice over IP
COMP9333 Advance Computer Networks Mini Conference QoS issues in Voice over IP Student ID: 3058224 Student ID: 3043237 Student ID: 3036281 Student ID: 3025715 QoS issues in Voice over IP Abstract: This
More informationLow-rate TCP-targeted Denial of Service Attack Defense
Low-rate TCP-targeted Denial of Service Attack Defense Johnny Tsao Petros Efstathopoulos University of California, Los Angeles, Computer Science Department Los Angeles, CA E-mail: {johnny5t, pefstath}@cs.ucla.edu
More informationTCP and Wireless Networks Classical Approaches Optimizations TCP for 2.5G/3G Systems. Lehrstuhl für Informatik 4 Kommunikation und verteilte Systeme
Chapter 2 Technical Basics: Layer 1 Methods for Medium Access: Layer 2 Chapter 3 Wireless Networks: Bluetooth, WLAN, WirelessMAN, WirelessWAN Mobile Networks: GSM, GPRS, UMTS Chapter 4 Mobility on the
More information15-441: Computer Networks Homework 2 Solution
5-44: omputer Networks Homework 2 Solution Assigned: September 25, 2002. Due: October 7, 2002 in class. In this homework you will test your understanding of the TP concepts taught in class including flow
More informationFrequently Asked Questions
Frequently Asked Questions 1. Q: What is the Network Data Tunnel? A: Network Data Tunnel (NDT) is a software-based solution that accelerates data transfer in point-to-point or point-to-multipoint network
More informationRecovery Modeling in MPLS Networks
Proceedings of the Int. Conf. on Computer and Communication Engineering, ICCCE 06 Vol. I, 9-11 May 2006, Kuala Lumpur, Malaysia Recovery Modeling in MPLS Networks Wajdi Al-Khateeb 1, Sufyan Al-Irhayim
More informationData Networks Summer 2007 Homework #3
Data Networks Summer Homework # Assigned June 8, Due June in class Name: Email: Student ID: Problem Total Points Problem ( points) Host A is transferring a file of size L to host B using a TCP connection.
More informationA Survey on Congestion Control Mechanisms for Performance Improvement of TCP
A Survey on Congestion Control Mechanisms for Performance Improvement of TCP Shital N. Karande Department of Computer Science Engineering, VIT, Pune, Maharashtra, India Sanjesh S. Pawale Department of
More informationTCP in Wireless Mobile Networks
TCP in Wireless Mobile Networks 1 Outline Introduction to transport layer Introduction to TCP (Internet) congestion control Congestion control in wireless networks 2 Transport Layer v.s. Network Layer
More informationTHIS PAPER describes a new framework the allocatedcapacity
362 IEEE/ACM TRANSACTIONS ON NETWORKING, VOL. 6, NO. 4, AUGUST 1998 Explicit Allocation of Best-Effort Packet Delivery Service David D. Clark, Fellow, IEEE, and Wenjia Fang Abstract This paper presents
More informationPerformance of networks containing both MaxNet and SumNet links
Performance of networks containing both MaxNet and SumNet links Lachlan L. H. Andrew and Bartek P. Wydrowski Abstract Both MaxNet and SumNet are distributed congestion control architectures suitable for
More informationA Passive Method for Estimating End-to-End TCP Packet Loss
A Passive Method for Estimating End-to-End TCP Packet Loss Peter Benko and Andras Veres Traffic Analysis and Network Performance Laboratory, Ericsson Research, Budapest, Hungary {Peter.Benko, Andras.Veres}@eth.ericsson.se
More informationMobile Communications Chapter 9: Mobile Transport Layer
Mobile Communications Chapter 9: Mobile Transport Layer Motivation TCP-mechanisms Classical approaches Indirect TCP Snooping TCP Mobile TCP PEPs in general Additional optimizations Fast retransmit/recovery
More informationApplying Active Queue Management to Link Layer Buffers for Real-time Traffic over Third Generation Wireless Networks
Applying Active Queue Management to Link Layer Buffers for Real-time Traffic over Third Generation Wireless Networks Jian Chen and Victor C.M. Leung Department of Electrical and Computer Engineering The
More informationClearing the Way for VoIP
Gen2 Ventures White Paper Clearing the Way for VoIP An Alternative to Expensive WAN Upgrades Executive Overview Enterprises have traditionally maintained separate networks for their voice and data traffic.
More informationQuality of Service using Traffic Engineering over MPLS: An Analysis. Praveen Bhaniramka, Wei Sun, Raj Jain
Praveen Bhaniramka, Wei Sun, Raj Jain Department of Computer and Information Science The Ohio State University 201 Neil Ave, DL39 Columbus, OH 43210 USA Telephone Number: +1 614-292-3989 FAX number: +1
More informationIntroduction to Quality of Service. Andrea Bianco Telecommunication Network Group firstname.lastname@polito.it http://www.telematica.polito.
Introduction to Quality of Service Andrea Bianco Telecommunication Network Group firstname.lastname@polito.it http://www.telematica.polito.it/ QoS Issues in Telecommunication Networks - 1 Quality of service
More informationCommunication Networks. MAP-TELE 2011/12 José Ruela
Communication Networks MAP-TELE 2011/12 José Ruela Network basic mechanisms Introduction to Communications Networks Communications networks Communications networks are used to transport information (data)
More informationEffects of Filler Traffic In IP Networks. Adam Feldman April 5, 2001 Master s Project
Effects of Filler Traffic In IP Networks Adam Feldman April 5, 2001 Master s Project Abstract On the Internet, there is a well-documented requirement that much more bandwidth be available than is used
More informationSJBIT, Bangalore, KARNATAKA
A Comparison of the TCP Variants Performance over different Routing Protocols on Mobile Ad Hoc Networks S. R. Biradar 1, Subir Kumar Sarkar 2, Puttamadappa C 3 1 Sikkim Manipal Institute of Technology,
More informationNetworking part 3: the transport layer
Networking part 3: the transport layer Juliusz Chroboczek Université de Paris-Diderot (Paris 7) September 2011 Summary of the previous episodes Episode 1: switching, packet switching and the Internet.
More informationSupporting VoIP in IEEE802.11 Distributed WLANs
Supporting VoIP in IEEE802.11 Distributed WLANs Zuo Liu Supervisor: Dr. Nick Filer July 2012 1 Voice VoIP Applications Constant Streaming Traffic Packetize interval usually 10-30 ms 8 160 bytes each packet
More informationTCP for Wireless Networks
TCP for Wireless Networks Outline Motivation TCP mechanisms Indirect TCP Snooping TCP Mobile TCP Fast retransmit/recovery Transmission freezing Selective retransmission Transaction oriented TCP Adapted
More informationPART III. OPS-based wide area networks
PART III OPS-based wide area networks Chapter 7 Introduction to the OPS-based wide area network 7.1 State-of-the-art In this thesis, we consider the general switch architecture with full connectivity
More informationComputer Network. Interconnected collection of autonomous computers that are able to exchange information
Introduction Computer Network. Interconnected collection of autonomous computers that are able to exchange information No master/slave relationship between the computers in the network Data Communications.
More informationOutline. TCP connection setup/data transfer. 15-441 Computer Networking. TCP Reliability. Congestion sources and collapse. Congestion control basics
Outline 15-441 Computer Networking Lecture 8 TCP & Congestion Control TCP connection setup/data transfer TCP Reliability Congestion sources and collapse Congestion control basics Lecture 8: 09-23-2002
More informationHigh-Speed TCP Performance Characterization under Various Operating Systems
High-Speed TCP Performance Characterization under Various Operating Systems Y. Iwanaga, K. Kumazoe, D. Cavendish, M.Tsuru and Y. Oie Kyushu Institute of Technology 68-4, Kawazu, Iizuka-shi, Fukuoka, 82-852,
More informationOptimization of Communication Systems Lecture 6: Internet TCP Congestion Control
Optimization of Communication Systems Lecture 6: Internet TCP Congestion Control Professor M. Chiang Electrical Engineering Department, Princeton University ELE539A February 21, 2007 Lecture Outline TCP
More informationProtagonist International Journal of Management And Technology (PIJMT) Online ISSN- 2394-3742. Vol 2 No 3 (May-2015) Active Queue Management
Protagonist International Journal of Management And Technology (PIJMT) Online ISSN- 2394-3742 Vol 2 No 3 (May-2015) Active Queue Management For Transmission Congestion control Manu Yadav M.Tech Student
More informationAnalysis and Detection of a Denial-of-Service Attack Scenario generated by TCP Receivers to Edge Network
Analysis and Detection of a Denial-of-Service Attack Scenario generated by TCP Receivers to Edge Network V. Anil Kumar 1 and Dorgham Sisalem 2 (anil@cmmacs.ernet.in, sisalem@fokus.fhg.de) 1 CSIR Centre
More informationJoint ITU-T/IEEE Workshop on Carrier-class Ethernet
Joint ITU-T/IEEE Workshop on Carrier-class Ethernet Quality of Service for unbounded data streams Reactive Congestion Management (proposals considered in IEE802.1Qau) Hugh Barrass (Cisco) 1 IEEE 802.1Qau
More informationAdopting SCTP and MPLS-TE Mechanism in VoIP Architecture for Fault Recovery and Resource Allocation
Adopting SCTP and MPLS-TE Mechanism in VoIP Architecture for Fault Recovery and Resource Allocation Fu-Min Chang #1, I-Ping Hsieh 2, Shang-Juh Kao 3 # Department of Finance, Chaoyang University of Technology
More informationA Survey: High Speed TCP Variants in Wireless Networks
ISSN: 2321-7782 (Online) Volume 1, Issue 7, December 2013 International Journal of Advance Research in Computer Science and Management Studies Research Paper Available online at: www.ijarcsms.com A Survey:
More informationCROSS LAYER BASED MULTIPATH ROUTING FOR LOAD BALANCING
CHAPTER 6 CROSS LAYER BASED MULTIPATH ROUTING FOR LOAD BALANCING 6.1 INTRODUCTION The technical challenges in WMNs are load balancing, optimal routing, fairness, network auto-configuration and mobility
More information51-20-99 Traffic Control Functions in ATM Networks Byung G. Kim Payoff
51-20-99 Traffic Control Functions in ATM Networks Byung G. Kim Payoff A standard monitoring algorithm for two traffic types--constant bit rate (e.g., voice) and variable bit rate (e.g., compressed video)--can
More informationRouter-assisted congestion control. Lecture 8 CS 653, Fall 2010
Router-assisted congestion control Lecture 8 CS 653, Fall 2010 TCP congestion control performs poorly as bandwidth or delay increases Shown analytically in [Low01] and via simulations Avg. TCP Utilization
More informationCOMP 361 Computer Communications Networks. Fall Semester 2003. Midterm Examination
COMP 361 Computer Communications Networks Fall Semester 2003 Midterm Examination Date: October 23, 2003, Time 18:30pm --19:50pm Name: Student ID: Email: Instructions: 1. This is a closed book exam 2. This
More informationAn Improved TCP Congestion Control Algorithm for Wireless Networks
An Improved TCP Congestion Control Algorithm for Wireless Networks Ahmed Khurshid Department of Computer Science University of Illinois at Urbana-Champaign Illinois, USA khurshi1@illinois.edu Md. Humayun
More informationEXAMPLES AND PROBLEMS. Competence Based Education Internet Protocols
EXAMPLES AND PROBLEMS Competence Based Education Internet Protocols Example 1 In following figure frames are generated at node A and sent to node C through node B. Determine the minimum transmission rate
More informationCongestions and Control Mechanisms n Wired and Wireless Networks
International OPEN ACCESS Journal ISSN: 2249-6645 Of Modern Engineering Research (IJMER) Congestions and Control Mechanisms n Wired and Wireless Networks MD Gulzar 1, B Mahender 2, Mr.B.Buchibabu 3 1 (Asst
More information4 Internet QoS Management
4 Internet QoS Management Rolf Stadler School of Electrical Engineering KTH Royal Institute of Technology stadler@ee.kth.se September 2008 Overview Network Management Performance Mgt QoS Mgt Resource Control
More informationAdaptive Sampling for Network Performance Measurement Under Voice Traffic
Sampling for Network Performance Measurement Under Voice Traffic Wenhong Ma and Changcheng Huang Optical Networks Laboratory Department of Systems and Computer Engineering, Carleton University 1125 Colonel
More informationComputer Networks. Chapter 5 Transport Protocols
Computer Networks Chapter 5 Transport Protocols Transport Protocol Provides end-to-end transport Hides the network details Transport protocol or service (TS) offers: Different types of services QoS Data
More informationIP Traffic Engineering over OMP technique
IP Traffic Engineering over OMP technique 1 Károly Farkas, 1 Zoltán Balogh, 2 Henrik Villför 1 High Speed Networks Laboratory Department of Telecommunications and Telematics Technical University of Budapest,
More informationImproving our Evaluation of Transport Protocols. Sally Floyd Hamilton Institute July 29, 2005
Improving our Evaluation of Transport Protocols Sally Floyd Hamilton Institute July 29, 2005 Computer System Performance Modeling and Durable Nonsense A disconcertingly large portion of the literature
More informationTCP and UDP Performance for Internet over Optical Packet-Switched Networks
TCP and UDP Performance for Internet over Optical Packet-Switched Networks Jingyi He S-H Gary Chan Department of Electrical and Electronic Engineering Department of Computer Science Hong Kong University
More informationComparative Analysis of Congestion Control Algorithms Using ns-2
www.ijcsi.org 89 Comparative Analysis of Congestion Control Algorithms Using ns-2 Sanjeev Patel 1, P. K. Gupta 2, Arjun Garg 3, Prateek Mehrotra 4 and Manish Chhabra 5 1 Deptt. of Computer Sc. & Engg,
More informationTCP Trunking for Bandwidth Management of Aggregate Traffic
TCP Trunking for Bandwidth Management of Aggregate Traffic H.T. Kung kung@eecs.harvard.edu Division of Engineering and Applied Sciences Harvard University Cambridge, MA 02138, USA S.Y. Wang shieyuan@csie.nctu.edu.tw
More informationAnalysis of IP Network for different Quality of Service
2009 International Symposium on Computing, Communication, and Control (ISCCC 2009) Proc.of CSIT vol.1 (2011) (2011) IACSIT Press, Singapore Analysis of IP Network for different Quality of Service Ajith
More informationQoS Parameters. Quality of Service in the Internet. Traffic Shaping: Congestion Control. Keeping the QoS
Quality of Service in the Internet Problem today: IP is packet switched, therefore no guarantees on a transmission is given (throughput, transmission delay, ): the Internet transmits data Best Effort But:
More informationNetwork management and QoS provisioning - QoS in the Internet
QoS in the Internet Inernet approach is based on datagram service (best effort), so provide QoS was not a purpose for developers. Mainly problems are:. recognizing flows;. manage the issue that packets
More informationSimulation-Based Comparisons of Solutions for TCP Packet Reordering in Wireless Network
Simulation-Based Comparisons of Solutions for TCP Packet Reordering in Wireless Network 作 者 :Daiqin Yang, Ka-Cheong Leung, and Victor O. K. Li 出 處 :Wireless Communications and Networking Conference, 2007.WCNC
More information1 All authors contributed equally to this paper and their names are listed in no particular order.
A Comparative Evaluation of Internet Pricing Models: Smart Markets and Dynamic Capacity Contracting Ranjitha Singh, Murat Yuksel, Shivkumar Kalyanaraman, T. Ravichandran 1 Rensselaer Polytechnic Institute,
More informationAnalysis of traffic engineering parameters while using multi-protocol label switching (MPLS) and traditional IP networks
Analysis of traffic engineering parameters while using multi-protocol label switching (MPLS) and traditional IP networks Faiz Ahmed Electronic Engineering Institute of Communication Technologies, PTCL
More informationPerformance measurements of STANAG 5066 and Applications Running over STANAG 5066
Performance measurements of STANAG 5066 and Applications Running over STANAG 5066 Steve Kille CEO September 14 th 2009 Why Measure? The End Goal is to run Applications and Services over HF Radio Maximize
More information4 High-speed Transmission and Interoperability
4 High-speed Transmission and Interoperability Technology 4-1 Transport Protocols for Fast Long-Distance Networks: Comparison of Their Performances in JGN KUMAZOE Kazumi, KOUYAMA Katsushi, HORI Yoshiaki,
More informationInternet Traffic Management
Mälardalen University Licentiate Thesis No.95 Internet Traffic Management Henrik Abrahamsson November 2008 School of Innovation, Design and Engineering Mälardalen University Västerås, Sweden Copyright
More informationRandom Early Detection Gateways for Congestion Avoidance
Random Early Detection Gateways for Congestion Avoidance Sally Floyd and Van Jacobson Lawrence Berkeley Laboratory University of California floyd@eelblgov van@eelblgov To appear in the August 1993 IEEE/ACM
More informationImproved Multiple File Transfer Protocol using Extended features of SCTP
Vol.2, Issue.3, May-June 2012 pp-1034-1038 ISSN: 2249-6645 Improved Multiple File Transfer Protocol using Extended features of SCTP Prabhanshu Jaiswal 1,Gaurav Agrawal 2,Pushpendra Singh 3, Dr. A.K. Sharma
More informationTCP/IP Over Lossy Links - TCP SACK without Congestion Control
Wireless Random Packet Networking, Part II: TCP/IP Over Lossy Links - TCP SACK without Congestion Control Roland Kempter The University of Alberta, June 17 th, 2004 Department of Electrical And Computer
More informationVoIP Network Dimensioning using Delay and Loss Bounds for Voice and Data Applications
VoIP Network Dimensioning using Delay and Loss Bounds for Voice and Data Applications Veselin Rakocevic School of Engineering and Mathematical Sciences City University, London, UK V.Rakocevic@city.ac.uk
More informationApplication Level Congestion Control Enhancements in High BDP Networks. Anupama Sundaresan
Application Level Congestion Control Enhancements in High BDP Networks Anupama Sundaresan Organization Introduction Motivation Implementation Experiments and Results Conclusions 2 Developing a Grid service
More informationHow To Provide Qos Based Routing In The Internet
CHAPTER 2 QoS ROUTING AND ITS ROLE IN QOS PARADIGM 22 QoS ROUTING AND ITS ROLE IN QOS PARADIGM 2.1 INTRODUCTION As the main emphasis of the present research work is on achieving QoS in routing, hence this
More informationother. A B AP wired network
1 Routing and Channel Assignment in Multi-Channel Multi-Hop Wireless Networks with Single-NIC Devices Jungmin So + Nitin H. Vaidya Department of Computer Science +, Department of Electrical and Computer
More informationComputer Networks Homework 1
Computer Networks Homework 1 Reference Solution 1. (15%) Suppose users share a 1 Mbps link. Also suppose each user requires 100 kbps when transmitting, but each user transmits only 10 percent of the time.
More informationAN OVERVIEW OF QUALITY OF SERVICE COMPUTER NETWORK
Abstract AN OVERVIEW OF QUALITY OF SERVICE COMPUTER NETWORK Mrs. Amandeep Kaur, Assistant Professor, Department of Computer Application, Apeejay Institute of Management, Ramamandi, Jalandhar-144001, Punjab,
More informationD1.2 Network Load Balancing
D1. Network Load Balancing Ronald van der Pol, Freek Dijkstra, Igor Idziejczak, and Mark Meijerink SARA Computing and Networking Services, Science Park 11, 9 XG Amsterdam, The Netherlands June ronald.vanderpol@sara.nl,freek.dijkstra@sara.nl,
More informationAn Improvement Mechanism for Low Priority Traffic TCP Performance in Strict Priority Queueing
An Improvement Mechanism for Low Priority Traffic TCP Performance in Strict Priority Queueing Mahdi Rahimi and Vahab Pournaghshband Advanced Network and Security Research Laboratory Computer Science Department
More informationDynamic Congestion-Based Load Balanced Routing in Optical Burst-Switched Networks
Dynamic Congestion-Based Load Balanced Routing in Optical Burst-Switched Networks Guru P.V. Thodime, Vinod M. Vokkarane, and Jason P. Jue The University of Texas at Dallas, Richardson, TX 75083-0688 vgt015000,
More informationResearch of TCP ssthresh Dynamical Adjustment Algorithm Based on Available Bandwidth in Mixed Networks
Research of TCP ssthresh Dynamical Adjustment Algorithm Based on Available Bandwidth in Mixed Networks 1 Wang Zhanjie, 2 Zhang Yunyang 1, First Author Department of Computer Science,Dalian University of
More informationEXTENDING NETWORK KNOWLEDGE: MAKING OLSR A QUALITY OF SERVICE CONDUCIVE PROTOCOL
EXTENDING NETWORK KNOWLEDGE: MAKING OLSR A QUALITY OF SERVICE CONDUCIVE PROTOCOL by Pedro Eduardo Villanueva-Pena, Thomas Kunz Carleton University January, 2006 This report examines mechanisms to gradually
More informationPerformance Evaluation of AODV, OLSR Routing Protocol in VOIP Over Ad Hoc
(International Journal of Computer Science & Management Studies) Vol. 17, Issue 01 Performance Evaluation of AODV, OLSR Routing Protocol in VOIP Over Ad Hoc Dr. Khalid Hamid Bilal Khartoum, Sudan dr.khalidbilal@hotmail.com
More informationChapter 3 ATM and Multimedia Traffic
In the middle of the 1980, the telecommunications world started the design of a network technology that could act as a great unifier to support all digital services, including low-speed telephony and very
More informationProvisioning algorithm for minimum throughput assurance service in VPNs using nonlinear programming
Provisioning algorithm for minimum throughput assurance service in VPNs using nonlinear programming Masayoshi Shimamura (masayo-s@isnaistjp) Guraduate School of Information Science, Nara Institute of Science
More informationAdaptive DCF of MAC for VoIP services using IEEE 802.11 networks
Adaptive DCF of MAC for VoIP services using IEEE 802.11 networks 1 Mr. Praveen S Patil, 2 Mr. Rabinarayan Panda, 3 Mr. Sunil Kumar R D 1,2,3 Asst. Professor, Department of MCA, The Oxford College of Engineering,
More informationTCP Westwood for Wireless
TCP Westwood for Wireless מבוא רקע טכני בקרת עומס ב- TCP TCP על קשר אלחוטי שיפור תפוקה עם פרוטוקול TCP Westwood סיכום.1.2.3.4.5 Seminar in Computer Networks and Distributed Systems Hadassah College Spring
More informationQuality of Service in the Internet. QoS Parameters. Keeping the QoS. Traffic Shaping: Leaky Bucket Algorithm
Quality of Service in the Internet Problem today: IP is packet switched, therefore no guarantees on a transmission is given (throughput, transmission delay, ): the Internet transmits data Best Effort But:
More informationApplications. Network Application Performance Analysis. Laboratory. Objective. Overview
Laboratory 12 Applications Network Application Performance Analysis Objective The objective of this lab is to analyze the performance of an Internet application protocol and its relation to the underlying
More informationPer-Flow Queuing Allot's Approach to Bandwidth Management
White Paper Per-Flow Queuing Allot's Approach to Bandwidth Management Allot Communications, July 2006. All Rights Reserved. Table of Contents Executive Overview... 3 Understanding TCP/IP... 4 What is Bandwidth
More informationArchitecture and Performance of the Internet
SC250 Computer Networking I Architecture and Performance of the Internet Prof. Matthias Grossglauser School of Computer and Communication Sciences EPFL http://lcawww.epfl.ch 1 Today's Objectives Understanding
More informationTraffic Engineering & Network Planning Tool for MPLS Networks
Traffic Engineering & Network Planning Tool for MPLS Networks Dr. Associate Professor, Department of Electrical Engineering Indian Institute of Technology Bombay, Powai, Mumbai 76 Founder & Director, Vegayan
More informationBandwidth Measurement in Wireless Networks
Bandwidth Measurement in Wireless Networks Andreas Johnsson, Bob Melander, and Mats Björkman {andreas.johnsson, bob.melander, mats.bjorkman}@mdh.se The Department of Computer Science and Engineering Mälardalen
More information